Temporal analysis of canopy gap distribution in natural forests: comparing managed and unmanaged stands using Quickbird and UltraCam-D imagery

Aim of study: To investigate and analyze the temporal and spatial distribution patterns of canopy gaps in managed and unmanaged forest stands using remote sensing images from different sources, emphasizing the relevance of these patterns in the context of natural processes and human interventions. Area of study: Managed and unmanaged stands of Dr. Bahramnia Forestry Plan, Hyrcanian forest, Golestan Province, Iran. Material and methods: High-resolution satellite imagery, including QuickBird (2007) and UltraCam-D (2011) digital aerial images, were utilized to detect canopy gaps. This study employed pixel-based classification methods, such as Support Vector Machine (SVM) and Maximum Likelihood (ML) algorithms, along with object-based methods, including Nearest Neighbor (NN), Random Forest (RF), Decision Tree (DT), and Bayes, for classification and to extract canopy gaps. Spatial distribution patterns were analyzed using the Nearest Neighbor Index, spatial autocorrelation, and clustering based on gap size, employing global Getis-Ord and local Moran’s I statistics. Main results: The results indicated that pixel-based classification methods achieved higher accuracy in mapping gap and non-gap areas, with the ML algorithm attaining 98.763% accuracy and a Kappa coefficient of 0.974 in 2007, and the SVM algorithm achieving 98.899% accuracy and a Kappa coefficient of 0.976 in 2011. The Nearest Neighbor Index revealed a regular spatial distribution of gaps in both managed and unmanaged stands. Clustering analysis showed distinct patterns for small, medium, and large gaps, with managed stands exhibiting clustering and unmanaged stands displaying random and regular patterns. Research highlights: This study underscores the importance of high-resolution remote sensing data and advanced classification algorithms for accurately mapping canopy gap dynamics, providing valuable insights for sustainable forest management.